Development and Application of LIBS-XRF Coupled Multi-spectrum Coal Quality Analyser (Invited)
Tian Zhihui,Wang Shuqing,Zhang Lei,Zhang Peihua,Ye Zefu,Zhu Zhujun,Dong Lei,Ma Weiguang,Yin Wangbao,Xion Liantuan,Jia Suotang
DOI: https://doi.org/10.3788/gzxb20235203.0352109
IF: 0.6
2023-01-01
ACTA PHOTONICA SINICA
Abstract:Thermal power plants in China have the dual tasks of energy security and energy conservation and emission reduction,with coal accounting for 50% to 70% of their operating costs. Faced with the implementation of energy conservation and emission reduction,low-carbon environmental protection,and energy transformation policies,promoting the clean and efficient utilization of coal has become the primary task of thermal power plants. Therefore, measuring the coal quality, pricing according to the quality, and optimizing combustion are the important ways for their production and development. However, China has a large variety of coal and a large difference in coal quality, so thermal power plant generally has the problem that the actual supply of coal and boiler design do not match each other, results in high power generation costs and low combustion efficiency. In order to achieve the optimal control of coal blending and combustion in thermal power plant, the key is to achieve rapid quality analysis and fine management of incoming coal and fired coal. The common methods of coal quality analysis in power plants include manual assay, robotic assay, neutron activation, Laser-Induced Breakdown Spectroscopy (LIBS) and X-Ray Fluorescence Spectroscopy (XRF). Both manual and robotic assay use traditional national standard chemical analysis methods, but the former requires multiple equipments and is time-consuming, while the latter is bulky. In addition, although the analysis results of the traditional national standard method are reliable, it is difficult to analyze the coal in each vehicle or on the belt online due to its long time consumption, which cannot be used for accurate blending and optimized combustion control. Neutron activation online monitor is highly sensitive, but radioactive and expensive. LIBS has the advantages of fast online and simultaneous detection of multiple elements, but the measurement repeatability needs to be further improved. XRF has high repeatability, but it is unable to analyze organic light elements in coal. In this study, based on the proposed coupled multi-spectrum method of LIBS and XRF, we designed a new software-controlled rapid coal quality analyzer, which includes LIBS analysis module, XRF analysis module, sample feeding module, control module, and operation software. This analyzer not only plays the strengths of LIBS for full elemental analysis, but also inherits the advantages of XRF for high stability analysis, which can be used in power plants for fast and continuous analysis of coal pellets. In addition, the spectrum analysis based on Partial Least Squares regression(PLS) method was modeled for hundreds of coal samples. The analysis process of the spectral data included pre-processing of LIBS and XRF spectrum, coupled LIBS-XRF modeling, and model testing, where the accuracy of the model was characterized by the correlation coefficient(R-2)and the mean absolute error(Delta), and the repeatability was tested by the Standard Deviation (SD). The industrial testing and performance evaluation were also completed at Shanxi Sunshine Power Plant. We collected spectra of hundreds of coal samples and preprocessed them, then established prediction models using Partial Least Square(PLS)method, and finally completed industrial testing and performance evaluation in Shanxi Yangguang Power Plant. The test results showed that the R2 of the prediction models for calorific value, ash content, volatile matter, and sulfur content were 0.973, 0.986, 0.977, and 0. 979 respectively, and the average standard deviations were 0.11%, 0.49%, 0.15% and 0.09% respectively. The model results showed good accuracy and stability, and the measurement repeatability meets the requirements of national standards. The average absolute errors of the analyzer in predicting the calorific value, ash content, volatile matter and sulfur content of coal were 0.39 MJ/kg, 0.83%, 0.50% and 0.23% respectively, and the single measurement takes about 5.5 minutes, which can meet the needs of industrial practical application. This XRF-LIBS coal quality quantitative analysis technology with excellent measurement repeatability is expected to be applied to power plants, coking plants, coal washing plants, cement plants, coal chemical industry and other industrial fields that need to pay attention to coal quality at all times.